Felo API PlatformFelo API Platform
Back to models
Anthropicmessages

Claude Fable 5.2 API

Claude Fable 5.2 is Anthropic's Mythos-class multimodal model with a 1 million token context, built for long-horizon agentic work, coding, and research.

Claude Fable 5.2 is a Mythos-class model from Anthropic, a generation above the Opus line, and the publicly available, safeguarded version of the model shared with Claude Mythos 5.2. It accepts text and image input and returns text output, with a context window of 1 million tokens and an output limit of 128k tokens. It is available through the Claude API, Amazon Bedrock, Google Vertex AI, Microsoft Foundry, and Claude subscription plans. Data is not used for training and is deleted after 30 days in most cases.

The model is designed for long-horizon agentic work across multiple steps: it can plan, act, check, and adjust while keeping track of the goal. It produces outputs up to 30 percent faster than Claude Fable 5.1 and communicates more clearly than previous Fable and Mythos versions. Anthropic describes it as state-of-the-art on nearly all tested benchmarks of AI capability, with top or near-top placements in independent evaluations for coding, reasoning, and tool use. It is multilingual and shows substantially improved control of physical robots, including tasks it was not explicitly trained on.

It suits complex agentic workflows, research pipelines that gather and synthesize information from multiple sources, code generation and debugging across large codebases, customer-facing agents handling edge cases, document processing with classification, extraction, and structured output, orchestration of multi-agent systems, and high-stakes document processing. Reported strengths include solving two open problems in black hole physics during testing, building a working operating system, designing three functional electronic devices, and migrating a critical production service from Python to Rust internally, yielding a service that starts about 26 times faster. It is a covered model requiring mandatory 30-day data retention on all traffic, and flagged sensitive queries are redirected to a fallback model such as Claude Opus 4.8. Anthropic recommends Claude Opus 5.5 as the daily driver for most coding and knowledge tasks.

Limited timeFree trial
Open Playground ↗Create API Key ↗Getting started →

Updated:

MODEL ID

claude-fable-5-2
Context
1M
Max output
128K
Recommended API
messages
Input
text / image / file
Output
text

Highlights

  • 1 million token context window
  • Up to 128k token outputs
  • Text and image input
  • Up to 30 percent faster than Claude Fable 5.1
  • Long-horizon agentic planning and adjustment
  • Multilingual with safety classifiers

Providers

Best price
6 providers

Compare API prices across providers. Select a provider to view all charges and conditions.

Provider · Pricing details
Uptime
$8$40$0.2
$8$2$40$10$0.2$0.05
$8$3.2$40$16$0.2$0.08
$8$5.2$40$26$0.2$0.13
$8$40$0.2
$8$40$0.2

What $10 buys

Input
1.3M tokens
Output
250K tokens

Estimated from current input and output prices, before discounts.

Benchmarks

  • Intelligence Index

    Max, Default Fallback; ranked #1 of 56 models

    64.6

  • certain benchmark

    Benchmark name not specified in source

    64.9%

Scores as published by the model provider.

Reasoning

Default effort
high
Always enabled
Yes
Available effort levels
max / xhigh / high / medium / low

Quick Start

Drop-in code to call this model with Felo API Platform compatible APIs. Start in the Playground when you want to test prompts before writing code.

Open PlaygroundCreate API Key
1

Get your API key

Create a key from your Felo API Platform account and set it as an environment variable.

export FELO_API_KEY="YOUR_API_KEY"
2

Make your first request

Use this model ID with the Felo API Platform endpoint. Optional request headers such as X-Request-Id can help troubleshooting.

claude-fable-5-2

This model is reasoning-capable. If your selected protocol and model route expose reasoning controls, preserve the returned reasoning metadata when continuing a conversation.

const response = await fetch("https://openapi.felo.ai/api/v1/messages", {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.FELO_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
  "model": "claude-fable-5-2",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}),
});
if (!response.ok) throw new Error(await response.text());
console.log(await response.json());

Using third-party SDKs

Use compatible SDKs by setting baseURL to https://openapi.felo.ai/api/v1. Use the Anthropic-compatible base URL for Claude-style clients. https://openapi.felo.ai/api.

3

Enable streaming

Add stream: true to receive responses as server-sent events.

curl -N https://openapi.felo.ai/api/v1/messages \
  -H "Authorization: Bearer $FELO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "claude-fable-5-2",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "stream": true
}'

Endpoint

Felo API Platform exposes compatible model-calling surfaces. Pick the surface your app or agent client already supports.

SurfaceMethodURL

Messages

Anthropic-compatible messages for Claude-style clients.

POST

https://openapi.felo.ai/api/v1/messages

Parameters

NameTypeDefaultDescription
modelstringclaude-fable-5-2Public model ID to call.
max_tokensinteger—Maximum generated tokens.
messagesarray—Conversation messages for Chat Completions.
stop———
toolsarray—Tool calling definitions following compatible schema.
reasoning———
verbosity———
tool_choicestring or object—Controls whether a tool should be called.
response_formatobject—Request structured output when supported.
reasoning_effort———
include_reasoning———
structured_outputs———
max_completion_tokens———

Frequently asked questions

What is Claude Fable 5.2?

Claude Fable 5.2 is a Mythos-class model from Anthropic, a generation above the Opus line, and the publicly available, safeguarded version of the model shared with Claude Mythos 5.2. It accepts text and image input and returns text output, with a context window of 1 million tokens and an output limit of 128k tokens. It is available through the Claude API, Amazon Bedrock, Google Vertex AI, Microsoft Foundry, and Claude subscription plans. Data is not used for training and is deleted after 30 days in most cases. The model is designed for long-horizon agentic work across multiple steps: it can plan, act, check, and adjust while keeping track of the goal. It produces outputs up to 30 percent faster than Claude Fable 5.1 and communicates more clearly than previous Fable and Mythos versions. Anthropic describes it as state-of-the-art on nearly all tested benchmarks of AI capability, with top or near-top placements in independent evaluations for coding, reasoning, and tool use. It is multilingual and shows substantially improved control of physical robots, including tasks it was not explicitly trained on. It suits complex agentic workflows, research pipelines that gather and synthesize information from multiple sources, code generation and debugging across large codebases, customer-facing agents handling edge cases, document processing with classification, extraction, and structured output, orchestration of multi-agent systems, and high-stakes document processing. Reported strengths include solving two open problems in black hole physics during testing, building a working operating system, designing three functional electronic devices, and migrating a critical production service from Python to Rust internally, yielding a service that starts about 26 times faster. It is a covered model requiring mandatory 30-day data retention on all traffic, and flagged sensitive queries are redirected to a fallback model such as Claude Opus 4.8. Anthropic recommends Claude Opus 5.5 as the daily driver for most coding and knowledge tasks.

How is usage billed?

Your bill depends on the provider you use, its prices and your usage. Longer inputs or other conditions may have different prices.

What input modalities does Claude Fable 5.2 support?

It accepts text and image input and produces text output.

How large are the context window and output limit?

The context window is 1 million tokens and outputs can reach 128k tokens.

What tasks is Claude Fable 5.2 best suited for?

It is built for long-horizon agentic work, research pipelines, code generation and debugging, customer-facing agents, document processing, and multi-agent orchestration.

How does it differ from Claude Mythos 5.2?

Claude Fable 5.2 is the publicly available and safeguarded version of the model shared with Claude Mythos 5.2, which has safety restrictions lifted in approved domains and is restricted to vetted partners through Project Glasswing.

What happens to sensitive prompts?

Safety classifiers flag sensitive prompts in offensive cybersecurity, biology and life sciences, and chemistry, and flagged queries are automatically redirected to a fallback model such as Claude Opus 4.8.

Related models

Claude Fable 5.2 (batch)

Anthropicclaude-fable-5.2:batch

Anthropic's most capable generally available model, built for long-horizon agentic engineering, scientific research and large-scale batch processing with a 1 million token context.

Open-ended coding projectsScientific researchQuantitative finance analysisLong-horizon agentic work

Context

1M

Max output

128K

Input tokens

$4 / 1M tokens

Output tokens

$20 / 1M tokens

Claude Haiku 5.5

Anthropicclaude-haiku-5-5

Claude Haiku 5.5 is Anthropic's smallest, fastest and cheapest Claude 5.5 model, offering a 1,000,000-token context window, image and PDF input, and an adjustable effort setting.

Live customer supportClassification requestsSummarizationDatabase queries

Context

1M

Max output

128K

Input tokens · Base rate

$0.1 / 1M tokens

Output tokens · Base rate

$0.5 / 1M tokens

Claude Sonnet 5.5

Anthropicclaude-sonnet-5-5

Claude Sonnet 5.5 is Anthropic's fast, low-cost workhorse model in the Claude 5.5 family, offering a 1,000,000-token context window, image input, and benchmark results close to Claude Opus 5.5.

Scoped agentic tasksBug fixesDocument creationPresentation and deck building

Context

1M

Max output

128K

Input tokens

$2 / 1M tokens

Output tokens

$10 / 1M tokens

Claude Opus 5.5

Anthropicclaude-opus-5-5

Anthropic's Claude Opus 5.5 tops the Artificial Analysis Intelligence Index across all five reasoning effort levels, from low through max.

Complex multi-step reasoningDeep analytical workTop-tier intelligence tasksAdjustable-effort workloads

Context

1M

Max output

128K

Input tokens

$4 / 1M tokens

Output tokens

$20 / 1M tokens

Claude Fable 5.1

Anthropicclaude-fable-5-1

Claude Fable 5.1 is Anthropic's highest-tier model, offering a 1-million-token context, 128K-token output, and state-of-the-art coding and knowledge work.

Demanding reasoning tasksLong-horizon agentic workScientific discoveryEnterprise document workflows

Context

1M

Max output

128K

Input tokens

$10 / 1M tokens

Output tokens

$50 / 1M tokens

Claude Haiku 4.5

Anthropicclaude-haiku-4-5

Claude Haiku 4.5 is Anthropic's fastest, most cost-efficient Claude model, released October 15, 2025, with extended thinking, computer use, and strong coding performance.

Fast iteration and focused codingMulti-step tool use across filesCoding and well-defined subagent tasksCustomer support and real-time chat

Context

200K

Max output

64K

Input tokens

$1 / 1M tokens

Output tokens

$5 / 1M tokens